Extracellular potentials recorded from the brain may hold the key to a greater understanding of how the brain functions. By presenting high-resolution information concerning neural activity, they may also prove to be the first step in a pathway to restoration of motor function and enhanced communication for individuals with nerve damage. Typically, systems that obtain recordings of neural signals involve a percutaneous (""""""""through the skin"""""""") connection between electrodes and wires, with the wires generally traveling to various pieces of electronic equipment. This method of acquiring neural signals is accompanied by limitations and possible complications that can be undesirable within a research setting and are definitely problematic within the context of a human brain-machine interface (BMI). The chronic break in the skin increases the likelihood of infection while the tethering effect of the wires provides an unnatural restraint on motion. In order to address these concerns and limitations, a completely implantable 96-channel neural data acquisition system is being developed. An implanted system, however, presents its own set of difficulties. The proposed research addresses the problem of getting data from the implanted system out of the body. An implanted system removes the physical link provided by wires between the electrodes and external equipment. In the proposed system, this link is replaced by a wireless, radio frequency (RF) link. Due to the sheer quantity of data acquired by 96 channels and the limitations of the RF link, only a small fraction of the acquired data can actually be sent out of the body. Thus, a data reduction scheme that drastically cuts down the amount of data to be sent out of the body while preserving the information of interest and importance is needed. The proposed research aims to design, implement, and test 1) a data reduction scheme suitable for both the research environment and a BMI application; 2) a bidirectional telemetry link for sending data out of the body and providing commands and configuration information to the implanted portion of the system; and 3) further processing of neural data within the system with the goal of performing real-time spike sorting. Relevance to public health: A brain-machine interface (BMI) can provide a means of communication or a means for restoration of motor function for individuals with severe nerve damage due to spinal cord injury, stroke, or neurodegenerative diseases. In order for a BMI to be clinically viable, neural signals must be obtained without requiring a chronic break in the skin or significant restraints. This project addresses one of the major challenges facing the development of a high-channel count fully implantable neural data acquisition system, which is a critical component for a successful, clinically viable BMI. ? ? ?

Agency
National Institute of Health (NIH)
Institute
National Institute of Biomedical Imaging and Bioengineering (NIBIB)
Type
Predoctoral Individual National Research Service Award (F31)
Project #
5F31EB007897-02
Application #
7430461
Study Section
Special Emphasis Panel (ZRG1-F15-N (20))
Program Officer
Erim, Zeynep
Project Start
2007-06-01
Project End
2008-12-31
Budget Start
2008-06-01
Budget End
2008-12-31
Support Year
2
Fiscal Year
2008
Total Cost
$15,212
Indirect Cost
Name
Duke University
Department
Biomedical Engineering
Type
Schools of Engineering
DUNS #
044387793
City
Durham
State
NC
Country
United States
Zip Code
27705
Rizk, Michael; Bossetti, Chad A; Jochum, Thomas A et al. (2009) A fully implantable 96-channel neural data acquisition system. J Neural Eng 6:026002
Rizk, Michael; Wolf, Patrick D (2009) Optimizing the automatic selection of spike detection thresholds using a multiple of the noise level. Med Biol Eng Comput 47:955-66